93-94 truck VAF to MAF conversion project (1 Viewer)

[semlin]

Ok this is a spinoff from Landtank's MAF conversion thread. I had given up on this as an option and gone back to my obd2 conversion idea but recent events have made this option look feasible.

I have been chatting with Landtank for a while about adapting his MAF system to plug and play replace the 93-94 VAF system. The possible advantages would be a smoother throttle response, smoother idle (dispensing with the FPR vacuum assist), better airflow control and perhaps overcoming the tendency of 93-94s to run lean at w-o-t, especially with a supercharger and, in a perfect world, mileage and horsepower gains from better airflow along the lines that Landtank is projecting for his system. Basically, if it works, you are replacing a 1992 technology mechanical vane system with a 2005 hotwire system. The only limitation would be how responsive the older obd1 ecu, o2 sensors, injectors and TPS would be to the newer technology. That's what you call a "known unknown". I am optimistic based on the fact the obd1 and obd2 ecus in the old supras are basically plug and play (they both use a karman vortex air flow meter). The only significant difference I know of is that the obd1 system has left and right trim on the injectors each controlled by the two upstream o2 sensors.

EricE, who will hopefully chime in, also has a 93-94, and is game to help and make a voltage converter. Which is good because I had pretty much given up on this without him.

the basic technical problems are:

1. according to the FSM the voltage signal on a VAF is 5 to 0v not 0 to 5v as with the MAF

2. we don't know for sure the signal is linear or might have a curve that differs between systems. We strongly suspect it is linear.

3. we don't know if the voltage to airflow scale is the same for the vaf and maf. we do want to test this as best we can.

4. we don't know if the obd1 ecu is dumbed down for low airflow signals because the vaf is not reliable at low airf flow such as idle (the vacuum line on the 93 FPR compensates for this deficiency). Again, we want to test this.

the plan is:

1. verify the voltage signal at various common rpm for a VAF and MAF truck to see if the scale is similar or not. If the scale is different a MAF translator inline between the MAF and voltage converter will be needed (we haven't yet looked over the FSM but any suggestions on how to get that voltage reading easily are appreciated)

2. fabricate a circuit board to convert the voltage signal from a MAF from 0-5v to 5-0v so the OBD1 ECU can read it. This will plug inline between the MAF and the ECU.

3. testbed a stock 97 MAF on a 93 truck. Luckily I have an early 93 ECU spare with the 83,84,85 code that can be fried.

4. if that works look to a variation on landtank's system that will bolt on a 93/94 and be plug and play. I say a variation because I believe there are variations in the OD of the VAF and MAF tubes that will impact on hose fitment.

At this point our main limiting factor is the lack of a participating 95-97 truck near to me in Vancouver or Eric in Seattle. If we had that, we could temporarily cannibilize the MAF to run the tests on the 93 and also figure out any mounting/connection issues we are going to face replacing the VAF with a MAF designed to fit the 95-97 system.

 

[EricE]

Good writeup.
We should also link to the original thread and work by Landtank.
https://forum.ih8mud.com/showthread.php?t=132012&page=1

We are just interested in taking his work and trying to extend it to the '93 and '94 trucks so we can use his newer tech MAF in our VAF trucks.

I have been researching the circuit design. So far I think it can be as simple as one op amp chip, 3 resistors, and 3 small capacitors. [Assuming the response of both MAF and VAF are linear enough].

Another limiting factor for me is that I need to wait till my truck is out of the shop before I can take any voltage readings off the existing VAF. It is getting a head gasket replacement.

 

[semlin]

no rush, I'm in moab next week.

 

[Bruneti]

Semlin,

I've been reading up on Landtank's project with great interest. Thought I would share some of my thoughts and ideas on this problem.

For your situation, at first it seemed to me that a flow bench would be the only way to directly compare the MAF to the VAF with any certainty. However, it's not likely that anybody has access to one.

Then this thought crossed my mind:
What if you put both the VAF and MAF on the same truck (in series) for purposes of data collection. If you did it on a OBDII truck, you could datalog the MAF signal voltage through the OBDII port like Landtank did. Then you would just need to supply the VAF with proper supply voltage and ground and datalog it's signal output (volts) at the same time. You should be able to tell where to overlay the curves since they were from the same runs and determine the realationship between the signal voltages. If you knew what you were doing, you could even establish the mathematical equation to convert from one to the other regardless of if they were linear slopes or not. This would not risk an ecu, and would tell you a great deal of what you need to know. It would just take a little research in the EWD to get the pinout of the VAF, and you may need to buy a repair plug & pins (should be pretty cheap) to make your own test harness for the VAF. I don't know what the supply voltage is to the VAF. If it's 12v then no problem, but if it's 5v you may need an old cell phone charger or something to convert to the proper voltage.

Even if one of the sensors was more restrictive than the other, since they are in series the system will only flow the air of the more restrictive one and they can still be directly compared. What do you think?

 

[semlin]

very interesting idea, but we'd need a 96-97 donor truck and an obd2 scanner and I'm not sure if the obd2 actually gives the signal voltage.

Alternately, we could reverse your suggestion and hook up a MAF in series to one of our 93s.

 

[EricE]

Jury rigging both MAF and VAF into one system at the same time is an excellent idea. One just has to supply some voltage and ground to the disconnected one and then read off outputs from both. Good idea.

 

[landtank]

here is what I'l propse we do to get a kind of bench mark for the two systems.

I can attach a meter to the signal line at my ECU. I can then monitor both that voltage and the rpms of my truck so we can plot out the voltages at different RPMs.

Taking a page from RavenTai's play book, I can set my laptop with Autoenginuity running and displaying OBDII data and the meter right next to it so a digital camera will snap shot both simaltaniously.

Drop those points into a MS graph program and we have our slope ands limits of the stock MAF system.

 

[semlin]

rick, if it's not too much trouble then that would be great before we try rigging the in series approach. If we get substantially identical inverted numbers to a 93 then we can just skip the more accurate inline approach.

 

[landtank]

I can say right now that the stock MAF won't be linear. I know this from all the data files I've been looking at. You can see this when looking at the Fuel trim values. They would rise up in mid range and then drop back down in the higher range. So basically there was a spot near middle that read lower than actual air flow.

The new MAF has a linear FT along the RPM range so it's more linear than the original.

Personally I don't think it will matter too much. If we basically get the idle voltage to match and then something close to the upper limit and let the middle find it's own. It was nice with OBDII because we could monitor fuel trim to see how accurately we were. I'm not sure how to do that with OBDI.

 

[semlin]

rick, if the obd1 ecu takes the voltage signal literally when ordering up fuel, andn uses o2 sensor feedback to trim for flaws, then I agree that matching the range/scale is good enough. if it has built in compensation based on known flaws in the VAF before processing the o2 sensor feedback we will have a problem and the trim will go out of spec. since it uses a vacuum in the FPR to compensate for the known flaw at idle i am hoping this is a sign they left the ecu code alone.

 

[landtank]

another thing will be the wiring. Both sensors use an IAT sensor with the same range and wiring. That is THA and a Ground

The air flow for a MAF unit is 12vdc, Ground and signal. We need to check these three against the VAF wiring.

 

[landtank]

I just read through the 93 VAF sensor on TIS again and this could be a little more than what we think.

Basically the MAF unit outputs a voltage proportional to the air flow.

The VAF unit is basically a variable resistor that drops the signal voltage down by reducing the resistance as air flow increases.

So the circuit would need to be able to do that.

EricE, your up!

 

[Rookie2]

Okay... not to disrupt your all's three-way here, but for those of us trying to follow along at home, where could one go to get a better base knowledge of WTF you all are talking about?? In particular MAF's, ECU's, O2's, FPR's and how these components interact to regulate fuel and air flow.

Thank you,
Rookie2

 

[Bruneti]

Landtank - Did the info on the VAF sensor state if it also uses a source voltage of 12v?

Your idea of grabbing snapshots of a voltmeter would work, and is super simple. If you were thinking of capturing RPM so that you could repeat the test for the other sensor, then I'm not sure that will work. If the two meters flow differently, then I don't think you can assume they are flowing the same air at the same rpm. I think putting the two meters in series solves that problem because you don't even have to know the flow rate or the RPM. They both have to flow the same air at that instant in time. Then you can grab a snapshot of two voltmeters and directly relate their voltages at the instant the snapshot is taken. You would know for sure exactly what both sensors read at that point on their respective curves with no guesswork. For a 0-5 volt scale, a few data points should then be enough to establish the basic curve. If they're not linear, it even sounds as if EricE has the electronics skills to build a simple translator.

 

[landtank]

Landtank - Did the info on the VAF sensor state if it also uses a source voltage of 12v?

I'm not sure. What I saw was signal names like VC and EG. I'll have to look again and search the schematics to see what they are.

 

[semlin]

VS and THA are the VAF related ECM terminals.

There are four terminals shown on the VAF

-THA is the ECM terminal for the IAT.
-VS is the ECM terminal for the VAF itself.
-E2 is an ECM terminal. Can't tell if it is a ground or power but you measure voltage on the other terminals across this one.

Finally, rhere is another terminal on the VAF labelled VC that the FSM does not explain. It connects to the VCC terminal on the ECM which is a TPS terminal.

the FSM gives the following voltage for the VAF to terminal E2 which I assume is a ground.

VS at idle = 1.2 - 2.4 v
VS at 3000 rpm = 0.8 to 1.3

Also,
VS with ignition switch on
-measuring plate closed = 3.5 to 4.5v
-measuring plate fully open = 0.2 to 0.5v

THA is measured at 20 degrees celsius at 0.5-3.4v

 

[semlin]

some extracts from the FSM that may help

 

[semlin]

and here are the others i could find.

 

[semlin]

and this is the IAT set up

 

[Bruneti]

My W.A.G. is that VCC is your supply voltage (also for the TPS sensor?), and E2 is ground.